Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation

Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation Cristina Andrés-Iglesias a , Juan García-Serna b , Olimpio Montero c , Carlos A. Blanco a, ⁎ a Departamento de Ingeniería Agrícola y Forestal (Área de Tecnología de los Alimentos), ETS Ingenierías Agrarias (Universidad de Valladolid), Avda. de Madrid 44, 34004 Palencia, Spain b High Pressure Processes Group, Department of Chemical Engineering and Environmental Tech., University of Valladolid, 47011 Valladolid, Spain c Centre for Biotechnology Development (CDB), CSIC. Boecillo Technological Park, Av. Francisco Vallés 8, 47151 Boecillo, Valladolid, Spain abstractarticle info Article history: Received 27 May 2015 Received in revised form 9 July 2015 Accepted 12 July 2015 Available online xxxx Keywords: Alcohol-free beer Aspen HYSYS simulation Dealcoholization Volatile compounds Flavor perception HS-SPME The coupled operation of vacuum distillation process to produce alcohol free beer at laboratory scale and Aspen HYSYS simulation software was studied to define the chemical changes during the dealcoholization process in the aroma profiles of 2 different lager beers. At the lab-scale process, 2 different parameters were chosen to dealcoholize beer samples, 102 mbar at 50 °C and 200 mbar at 67 °C. Samples taken at different steps of the process were analyzed by HS-SPME–GC–MS focusing on the concentration of 7 flavor compounds, 5 alcohols and 2 esters. For simulation process, the EoS parameters of the Wilson-2 property package were adjusted to the experimental data and one more pressure was tested (60 mbar). Simulation methods represent a viable alternative to predict results of the volatile compound composition of a final dealcoholized beer. © 2015 Published by Elsevier Ltd. 1. Introduction The market of non-alcoholic brews has experienced a significant improvement during the past years that becomes motivated mainly by highly competitive markets, driving/drinking rules, health conditions incompatible with alcohol consumption and/or religious reasons (Andrés-Iglesias, Montero, Sancho, & Blanco, 2014; Blanco, Andrés-Iglesias, & Montero, 2014; Catarino & Mendes, 2011). Similarly, it is well-known that beer has positive effects and a whole range of properties, such as no fat or cholesterol content, free sugar content, high antioxidant, magnesium and soluble fiber content (Brányik et al., 2012), plus it provides essential vitamins and minerals contributing to a healthy balanced diet (Andrés-Iglesias, Blanco, Blanco, & Montero, 2014; Bamforth, 2001). Beer aroma profile is made by many volatile organic compounds at very low concentration (ppm level), which are responsible for its unique flavor (Catarino, Mendes, Madeira, & Ferreira, 2007). Levels of different chemical compounds, such as alcohols, esters, aldehydes, ketones, organic acids and phenols, can be found on beer composition, giving a specificflavor that contributes to the overall organoleptic properties of the final beer (Karlsson & Trägårdh, 1997). Among them, esters and alcohols are the main groups of aroma compounds. Esters are responsible for the sweet and fruity flavors of beer, while alcohols confer it an alcoholic, fruity and immature flavor (Andrés-Iglesias, Blanco et al., 2014; Andrés-Iglesias, Montero et al., 2014; Catarino, Ferreira, & Mendes, 2009). In low-alcohol and/or alcohol-free beer production, the different techniques used have to be able to reach the maximum alcohol by volume (ABV) established by the different countries legal regulations. In the majority of EU countries beers withlow alcohol content are divided into alcohol-free beers (≤0.5% ABV) and low-alcohol beers (≤1.2% ABV). In Spain, alcohol free beers are divided in non-alcohol beers (≤1.0% ABV) and ‘0.0%’beers (≤0.1% ABV). However, in the United States there should not be alcohol present in alcohol-free beers, while 0.5% ABV corresponds to the upper limit of non-alcoholic beers or ‘near-beers’(Olmo, Blanco, Palacio, Prádanos, & Hernández, 2014). At present, there are several methods for low alcohol beer production (Blanco et al., 2014). The strategies can be divided into two main groups: biological and physical methods (Brányik et al., 2012; Montanari, Marconi, Mayer, & Fantozzi, 2009; Olmo et al., 2014). While physical methods withdraw the ethanol from a fermented beer, biological methods aim at controlling the alcohol production during the fermentation process (Zürcher, Jakob, & Back, 2005). Biological methods can be achieved by either restricting ethanol formation or shortening the fermentation process. Obtaining low alcohol content via interrupted fermentation is accompanied by low contents of aroma and flavor compounds, and their products are often Food Research International xxx (2015) xxx–xxx ⁎Corresponding author. E-mail address: [email protected] (C.A. Blanco). FRIN-05933; No of Pages 10 http://dx.doi.org/10.1016/j.foodres.2015.07.017 0963-9969/© 2015 Published by Elsevier Ltd. Contents lists available at ScienceDirect Food Research International journal homepage: www.elsevier.com/locate/foodres Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 characterized by worty off-flavors. They are usually performed using traditional brewery equipment and hence do not require additional investments (Brányik et al., 2012; Catarino & Mendes, 2011). Other processes to avoid these limitations include the use of special or immobilized yeasts as well as the use of low sugar raw materials (Catarino & Mendes, 2011; Pickering, 2000). The use of special yeasts for a low alcohol beer production process increases the costs with the need of yeast selection, or genetic modification of the production organisms. However, suitable selected yeasts can contribute significantly to the product sensorial quality improvement. Alcohol free beer production processes by continuous fermentation with immobilized yeast, whose objectiveis limited alcohol formation,andwhich requires special equipment and material. In this latter case, high investment costs are required but are justified by a higher productivity of the continuous processes. In general, producing alcohol-free beer by biological methods makes impossible theproduction ofalcohol-freebeers with alcoholcontent close to zero (Brányik et al., 2012). Physical methods require considerable investments into the special equipment for alcohol removal (Brányik et al., 2012). The most common separation processes used for beer dealcoholization are membranebased processes and heat treatment (Catarino et al., 2007). Membrane-based processes include reverse osmosis, nanofiltration, dialysis and pervaporation (Labanda, Vichi, Llorens, & López-Tamames, 2009). Heat treatment processes comprise evaporation and distillation; they are both under vacuum conditions to preserve the organoleptic properties by avoiding undesired secondary reactions (BelisarioSánchez, Taboada-Rodriguez, Marin-Iniesta, & López-Gómez, 2009). Furthermore, thermal processes to remove alcohol from regular beers can cause the loss of the original aroma (Blanco et al., 2014; Catarino et al., 2009) but their advantage is that they can remove ethanol from beers to levels close to zero (Brányik et al., 2012). Among these physical methods, for large scale dealcoholization, the vacuum evaporation is the most economic process (Zürcher et al., 2005). Distillation is a separation operation based on differences in volatility. If a mixture containing substances that differ in their volatility is brought to ebullition, the composition of the vapors released will be different from that of the boiling liquid. After condensation, the vapors constitute the “distillate”. The remaining liquid is called “residue” (Berk, 2013). The application of vacuum to distillation process enables to reduce the evaporation temperature and thus the thermal stress to beer (Zürcher et al., 2005). If the pressure is reduced, alcohol can be drawn off at much lower temperature (Brányik et al., 2012). Thermal processes to produce alcohol free beers are performed at temperatures between 30 and 60 °C at pressures of 60 to 200 mbar (Sohrabvandi, Mousavi, Razavi, Mortazavian, & Rezaei, 2010; Zürcher et al., 2005). The deterioration of beer quality by thermal dealcoholization depends mainly on the evaporation temperature and the period of exposure (Brányik et al., 2012). It is well known that most of the aroma compounds are lost in alcohol free beers during production by thermal processes. The aroma profile is clearly damaged and other, less pleasant flavors, like bready, worty or caramel notes can appear (Blanco et al., 2014; Catarino et al., 2009; Lehnert et al., 2009; Sohrabvandi et al., 2010). To compensate these disadvantages many breweries use a modified brewing technology for the production of a more aromatic original beer. Another attempt to compensate sensory disadvantages is by blendingdealcoholized beer with a small quantity of original beer or a beer aroma extract that can be recovered in evaporation plants with rectification columns. Since these attempts are not yet satisfactory further possibilities to improve the quality of these beers have been investigated (Zürcher et al., 2005). Owing to beer chemical compound characterization has improved notably nowadays, analysis of beer flavor compounds has been constantly optimized to obtain better results in relation to sensitivity and specificity (Andrés-Iglesias, Blanco et al., 2014; Andrés-Iglesias, Montero et al., 2014). Gas chromatography–mass spectrometry (GC–MS) is currently used to measure volatile compound concentrations in beer. Ethers, esters, acids, aldehydes, ketones, alcohols, sulfur compounds, hydrocarbon compounds, alicyclic compounds, heterocyclic compounds and aromatic compounds can be measured simultaneously by using GC–MS methods (Andrés-Iglesias, Blanco et al., 2014; Andrés-Iglesias, Montero et al., 2014). The combination of solid phase microextraction (SPME) with gas chromatography (GC) or gas chromatography–mass spectrometry (GC–MS) has proven to be a sensitive and precise method for the analysis of different classes of volatile compounds (Dong et al., 2013). Beer dealcoholization via vacuum distillation in a batch system can be assumed to resemble differential distillation at reduced pressure. The principles of differential distillation are well established since the beginning of chemical engineering knowledge. Thus, this type of distillation is often known as “Rayleigh distillation”. Lord Rayleigh's law is based on a dynamic material balance to the volatile compound of a two component mixture coupled to the global mass balance (Berk, 2013). Extendingthebalance to a multicomponent mixture wasstudied in deep by several authors such as Lang et al. (1994) and, Yatim et al. (1993) who modified the process for the addition of an extractive agent, or including sieves. An interesting comparative study was conducted by Zürcher et al. (2005) using lab scale batch and continuous distillation as well as an industrial scale plant. They investigated the beer dealcoholization at 60 and 150 mbar by following changes in a number of compounds, e.g., ethanol, 1-propanol, ethyl acetate, 2methylpropanol, 3-methylpropanol and several esters. However, they did not simulate the process. In addition, several authors have investigated the simulation of spirits production by this process. Claus and Berglund studied fruit brandy distillation using a batch column distillation. They simulated the process using CHEMCAD with good results using NRTL (NonRandom Two Liquids) equation of state (EoS) together with UNIFAC parameters (Claus & Berglund, 2005, 2009). On the other hand, Gaiser et al. simulated the whisky still distillation process using Aspen Plus software and selecting the NRTL-2 property package of that software; they claimed that this EoS provides a good approximation for ethanol–water azeotrope (Gaiser et al., 2002). Low alcohol and alcohol free beer consumption is increasing year by year, and often, these types of beverages are known to have a poor flavor profile in comparison to the original beer. In this sense, it becomes important to adjust the flavor of non-alcoholic beers to that of regular ones, which needs understanding of how the dealcoholization process modifies it, an issue of which the scientific info is scarce. In this work, we have combined lab scale differential vacuum distillation,aroma compoundanalysis and simulation of the dealcoholization process to shed light on this process. The main objective is to test a simulation environment that can explain the lab results, so that, it can be extrapolated to a similar process at industrial scale. For this, we have selected two model beers, one from Spain and one from Germany, and adjusted the interaction parameters of a thermodynamic model. To our knowledge, this is the first time that it is done for beers. 2. Material and methods 2.1. Samples and vacuum distillation dealcoholization experiments Two different big-scale lager beer brands were chosen for the study, one from Spain (S) and another one from Germany (G). Both of them were lager alcoholic beers containing 5.5 and 4.8% alcohol by volume (ABV) respectively, and were obtained as fresh as possible from the local market. Beer bottles were stored at 4 °C until dealcoholization process. 400 mL of beer were weighted and placed in 1 L flask of the vacuum distillation system for each experiment; the flask was covered with a black plastic material to avoid the light oxidation of the sample components. Subsequently, 10 μL of antifoam emulsion (E-900, AFCA) were added to reduce the foam and CO 2 content. 2C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 The experiments of beer dealcoholization by laboratory scale vacuum distillation were done at two different vacuum pressures and water bath temperatures. The temperature needed in the water bath is directly related to the total pressure by the phase equilibrium of the system, but has to be set to a slightly higher value to assure enough heat transfer. Thus, the first set of experiments was conducted at 102 mbar and 50 °C (corresponding to a saturation temperature of pure water, 46.2 °C) and the second set at 200 mbar and 67 °C (corresponding to a saturation temperature of pure water, 60.1 °C), A Rotavapor R-215 with vacuum pump V-700, vacuum controller V-850 and diagonal condenser (BÜCHI Labortechnik AG, Switzerland) was used. The flask rotation was fixed at 20 rpm and remained constant in all experiments. Each dealcoholization process was stopped at 15, 30, 45 and 60 min after the onset for sampling in order to analyze the different volatile compounds evaporated along with the ethanol. At the end of the distillation process, the residual beer was cooled in glass bottles and weighted for the material balance calculation. For all experiments the same steps were done. At the beginning of each experiment the water batch was refilleduntilthesamevolumeif necessary, once the batch reached the temperature, the experiment started at the rpm indicated above, the pressure was reached immediately and remained constant (±1) over the whole experiment and controlled by the vacuum controller. For the GC–MS analysis 15 mL dark vials sealed with PTFE–silicone septa (Supelco, USA) were used for sample preparation. Vials contained 2 g of NaCl (Scharlau, Scharlab S.L., Spain) and 5 mL of beer were added and stirred to solve the NaCl and homogenize the sample. A total of 60 samples were taken and analyzed from the original beers, and from residual beers at each time and dealcoholization process experiments. 2.2. Gas chromatography–mass spectrometry (GC–MS) equipment Volatile compounds were separated and detected by gas chromatography using an Agilent GC 6890N (Agilent Technologies, USA) equipped with an Agilent 5973 single quadrupole mass spectrometer (, Agilent Technologies, USA) as detector. A headspace solid phase microextraction (HS-SPME) manual equipment (Supelco, USA) was used for the extraction and concentration of the volatile compounds, which was carried out with 100 μm polydimethylsiloxan (PDMS) fiber (Sulpeco, USA). Prior to use, the SPME fiber was conditioned at 250 °C for 30 min in the GC injector, according to the manufacturer's instructions. Blank runs of the fiber were completed, before sampling, each day to ensure no carry-over of analytes. The chromatographic separations were accomplished using a BP-1 30 m × 0.32 mm × 1 μm capillary column (SGE Analytical Science, Australia). 2.3. Analysis of volatile compounds The volatile composition of beer samples wasmeasured by triplicate. Solid phase microextraction of compounds was performed at 30 °C for 45 min. The desorption was achieved in the injector of the GC chromatograph in splitless mode for 15 min, and the temperature was set at 250 °C as indicated by the manufacturer for PDMS fiber. Carrier gas was helium at a constant flow of 1.2 mL/min. The oven temperature was programmed as follows: initial temperature was set at 35 °C and kept for 7 min, this was followed by 2 ramps in which temperature was risen at 8 °C/min to 200 °C and kept this temperature for 5 min, and then temperature was risen at 10 °C/min to 250 °C, this temperature being kept for 10 min. The ionization energy was 70 eV, and detection and data acquisition were performed in scan mode from 37 to 350 Da. For identification data obtained in the GC–MS analysis were compared with m/z values compiled in the spectrum library WILEY. Validation of compound identification was carried out by comparison of MS spectra and retention times with those of commercialstandards. Quantification was carried out by using standard calibration curves of 2-methylbutanol (≥99.0%), 3-methylbutanol (≥99.0%), 2-phenylethanol (≥99.0%), ethyl acetate (≥99.5%), isobutanol (≥99.0%) (these from Sigma, USA). 1-Propanol ≥99.5% (Fluka, Sigma-Aldrich, USA) and isoamyl acetate ≥99.0%(Fisher,UK).Because1-propanolco-elutedwith ethanol, the extracted ion chromatogram (EIC) for the ion with m/z 60.05 and retention time of 3.10 min was used for quantification of this compound. 2.4. HYSYS simulation and parameters In order to simulate the system under study for the batch distillation of beer the following assumptions were considered: 1) The vacuum is done almost instantly and att = 0 thesystem is at the constant desired vacuum pressure. 2) Liquid composition is homogeneous and heat is uniformly distributed. 3) The flask has been simulated by a cylinder to simplify level calculation. 4) The heat flux for each data point is determined to match the time required for a certain vaporization volume. This is because the Rotavapor system can provide different heat flux depending on a number of variables (water level, flask location, ambient temperature, rotation speed, etc.). 5) No reaction occurs in the bulk liquid. The simulations have been carried out using HYSYS simulation software (Aspen inc. product) as it has a powerful non-steady state simulation tool. Wilson-2 property package waschosenin order to simulate the nonideal behavior of the liquid phase, while ideal gas is considered for the gas phase (as it was under reduced pressure conditions). The main simulation process flow diagram is depicted in Fig. 1.The main distillation vessel (V-100) has one feed stream-5 (virtual for simulation purposes set at almost zero flow), one heat source (Q-100), one liquid outlet stream-2 (virtual for simulation purposes set at almost zero flow) and one vapor outlet stream-3 (main distillation outlet). Themain calculations were carried outusing anExcel spreadsheet to determine the conversion between ppm and molar fraction values from experimental conditions to the simulation and vice versa. The main components simulated were: sucrose, ethanol, ethyl acetate, 1-propanol, isobutanol, isoamyl acetate, 2-methylbutanol, 3methylbutanol, 2-phenylethanol, water and nitrogen. Sucrose was used as a simulation trick to increment the density of water targeting the real value of 1010 kg/m 3 , for that purpose a concentration of 3% wt. was used in all simulation experiments. Nitrogen was used for simulation purposes mimicking the atmosphere of the Rotavapor. Initial values for compositions of the liquid were inserted in the “hold-up”values of the distillation vessel. The total pressure of stream-2 was fixed to the experimental absolute pressure, coinciding with the vessel initial pressure (i.e., 102 and 200 mbar). Fig. 1. HYSYS simulation model for a differential vacuum distillation. 3C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 As indicated in the assumptions, the heat flux was estimated to match the mass evaporated at each time sample point. This way, the simulation time is not as important as the evaporated mass, that is used as the x-axis variable as percentage of mass evaporated (%vapor). Thus, all experiments were carried out until 15, 30, 45 and 60 min, time when the dealcoholization process was stopped and the samples were collected. The % of vapor fraction (% Vf) was calculated as the percentageof initialmass of the beer minusthemassat the differentpoints of the simulation until the last mass (at 60 min of simulation) divided by the initial mass. Although the traditional ASTM D-86 curves for petroleum distillation are carried out in volume, in this case, mass was preferred to overcome density variations (ASTM-International, 2012). Furthermore, the heat flux could have varied along with the experiment. For this reason, we have considered this variable more accurate than experimental time itself. In addition to this, results could be transferred to a real vacuum distillation process with better scale-up chances. The developed software is available free in the web page of the research group of High Pressure Processes of the University of Valladolid (http://hpp.uva.es/software/) in the section for ‘Beer Distillation’. 3. Results and discussion Two lager beers were investigated in this study, one sample from Spain (S) and the other sample from Germany (G). Both samples were dealcoholized by vacuum distillation at laboratory scale at 2 different Fig. 2. Sample of TIC chromatogram for S beer sample, alcohol beer on the top and beer dealcoholized by laboratory vacuum distillation on the bottom. (1) 1-propanol, (2) ethyl acetate, (3) isobutanol, (4) 3-methylbutanol, (5) 2-methylbutanol, (6) isopentyl acetate, (7) 2-phenylethanol. Fig. 3. Ethanol behavior against the %vapor fraction on the left for S sample and for G sample on the right. 4C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 pressures and temperatures, 102 mbar, 50 °C and 200 mbar, 67 °C. A total of 45 compounds were identified, and 7 of them quantified by peak area. The profile of quantified volatiles consisted of 5 alcohols (1propanol, 2-methylpropanol, 2-methylbutanol, 3-methylbutanol and 2-phenylethanol) and 2 esters (ethyl acetate and isoamyl acetate). These compounds are considered as ones of the main alcohols and esters in beer (Blanco et al., 2014; Lehnert et al., 2009; Rodrigues, Caldeira, & Câmara, 2008;Willaert & Nedovic, 2006), the most abundant being ethyl acetate, isoamyl acetate, the amyl alcohols and isobutanol (Piddocke, Kreisz, Heldt-Hansen, Fog Nielsen, & Olsson, 2009). Also, these compounds are those currently analyzed by other authors because they are relevant flavor compounds in beer (Charry-Parra, DeJesús-Echevarria, & Perez, 2011; Kobayashi, Shimizu, & Shioya, 2008) and chosen as reference compounds when studying industrial processes for production of non-alcoholic beers (Catarino & Mendes, 2011; Mota et al., 2011). A typical total ion chromatogram (TIC) of a regular beer sample and its dealcoholized beer by laboratory scale vacuum distillation process is shown in Fig. 2. 3.1. Final ethanol content calculated by ASPEN HYSYS simulation During the differential distillation process, the most volatile fraction (ethanolic fraction) abandons the system in first place together with an increasing amount of water. In this work, we have focused on the analysis of the beer, rather than the evaporated volatile fraction (ethanolic fraction). Table 1 Percentageofthe vaporfractions(%Vf)of SandG samplesanditscorrespondent times,for both lab-scale vacuum distillation processes and the averages (%). Time, min 0 15 30 45 60 S 102 mbar 0 7.46 9.55 13.40 15.76 S 200 mbar 0 6.17 10.14 15.12 19.22 G 102 mbar 0 5.70 9.00 14.40 17.60 G 200 mbar 0 10.80 13.40 14.80 18.90 Average (% Vf) 0 7.53 10.52 14.43 17.87 Fig. 4. Ethanol concentration if the final volume is diluted or non-diluted at the two experimental pressures. Table 2 Retention time (Rt), concentration of volatile compounds (mg/L) delacoholized at 102 mbar, 50 °C and 200 mbar, 67 °C in the experiment (EXP), in simulations (SIM) and the standard deviation of the experimental value (St Dev). SIM EXP St Dev SIM EXP St Dev SIM EXP St Dev SIM EXP St Dev SIM EXP St Dev 102 mbar, 50 °C Rt S, 0 min S, 15 min S, 30 min S, 45 min S, 60 min 1-Propanol 3.10 5.56 5.56 2.47 1.22 2.09 0.36 0.50 0.25 0.12 0.05 0.24 0.00 0.01 0.00 0.00 Ethyl acetate 4.13 17.82 17.82 1.11 1.52 1.07 0.10 0.67 1.23 0.41 0.15 0.86 0.30 0.07 0.54 0.09 Isobutanol 4.60 9.41 9.41 2.52 5.67 4.97 2.03 4.38 4.68 2.73 2.35 3.67 2.56 1.63 3.26 2.10 3-Methylbutanol 9.11 40.99 40.99 2.73 25.01 17.44 1.45 19.37 17.15 7.14 10.43 11.51 1.54 7.24 7.33 1.05 2-Methylbutanol 9.27 13.37 13.37 0.94 7.78 5.26 0.54 5.83 5.29 2.14 2.91 3.55 0.63 1.91 1.93 0.27 Isopentyl acetate 14.09 1.92 1.92 0.33 0.36 0.11 0.08 0.15 0.10 0.09 0.02 0.09 0.08 0.00 0.09 0.08 2-Phenyl ethanol 19.82 34.01 34.01 1.39 37.07 40.46 6.26 37.95 48.76 17.74 39.51 62.04 9.92 40.16 85.28 3.90 102 mbar, 50 °C Rt G, 0 min G, 15 min G, 30 min G, 45 min G, 60 min 1-Propanol 3.10 8.93 8.93 3.48 2.36 6.99 2.20 0.52 0.28 0.01 0.02 0.17 0.03 0.00 0.00 0.00 Ethyl acetate 4.13 26.54 26.54 0.74 3.85 3.65 1.08 1.08 2.93 0.47 0.20 1.06 0.17 0.07 0.45 0.06 Isobutanol 4.60 10.47 10.47 2.66 6.89 6.88 2.08 4.48 4.32 2.21 2.07 2.88 1.93 1.12 2.87 1.93 3-Methylbutanol 9.11 43.77 43.77 2.12 28.46 27.85 2.79 18.58 13.69 2.02 8.58 4.92 0.82 4.65 5.13 0.88 2-Methylbutanol 9.27 11.54 11.54 0.42 7.27 7.48 1.18 4.50 3.49 0.70 1.86 1.12 0.43 0.90 1.22 0.25 Isopentyl acetate 14.09 2.58 2.58 0.25 0.67 0.27 0.15 0.15 0.16 0.10 0.01 0.13 0.09 0.00 0.11 0.08 2-Phenyl ethanol 19.82 37.69 37.69 4.68 40.00 53.95 5.40 41.46 56.92 5.71 43.17 69.57 6.81 43.71 75.17 4.58 200 mbar, 67 °C Rt S, 0 min S, 15 min S, 30 min S, 45 min S, 60 min 1-Propanol 3.10 5.56 5.56 0.69 3.63 3.58 0.90 1.51 2.51 0.94 0.30 0.00 0.00 0.02 0.00 0.00 Ethyl acetate 4.13 17.82 17.82 1.79 2.99 4.09 1.00 0.79 1.19 0.07 0.22 0.86 0.14 0.07 0.46 0.11 Isobutanol 4.60 9.41 9.41 2.87 5.60 4.88 1.83 3.06 4.25 1.65 1.40 3.67 1.71 0.53 3.03 1.47 3-Methylbutanol 9.11 40.99 40.99 4.05 26.58 20.44 4.01 15.70 16.90 0.83 7.88 11.51 2.29 3.34 7.35 0.70 2-Methylbutanol 9.27 13.37 13.37 1.57 8.40 6.44 1.18 4.70 5.18 0.67 2.17 3.55 0.75 0.80 2.20 0.29 Isopentyl acetate 14.09 1.92 1.92 0.10 0.63 0.20 0.07 0.17 0.12 0.04 0.02 0.09 0.05 0.00 0.07 0.04 2-Phenyl ethanol 19.82 34.01 34.01 2.59 36.63 43.36 7.14 38.28 53.41 3.85 39.66 62.04 6.35 40.17 70.65 3.61 200 mbar, 67 °C Rt G, 0 min G, 15 min G, 30 min G, 45 min G, 60 min 1-Propanol 3.10 8.93 8.93 3.32 1.11 0.40 0.11 0.26 0.30 0.02 0.22 0.14 0.02 0.01 0.00 0.00 Ethyl acetate 4.13 26.54 26.54 3.91 0.90 5.18 0.48 0.35 1.24 0.28 0.34 2.91 0.53 0.10 2.73 0.25 Isobutanol 4.60 10.47 10.47 3.29 2.43 5.07 1.76 1.22 3.46 1.56 1.24 1.93 1.53 0.36 1.82 1.34 3-Methylbutanol 9.11 43.77 43.77 6.60 12.11 19.95 0.57 6.50 9.48 1.31 6.64 5.04 0.88 2.18 2.94 0.76 2-Methylbutanol 9.27 11.54 11.54 1.17 2.85 4.85 0.50 1.43 2.29 0.28 1.45 1.17 0.34 0.40 0.64 0.22 Isopentyl acetate 14.09 2.58 2.58 0.37 0.11 0.49 0.05 0.02 0.16 0.06 0.02 0.09 0.05 0.00 0.08 0.07 2-Phenyl ethanol 19.82 37.69 37.69 5.40 42.27 54.97 2.93 43.47 62.58 10.36 43.27 59.41 8.68 43.98 59.97 1.01 5C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 Nevertheless, the concentration of ethanol in the ethanolic fraction in alcohol by volume percentage (% ABV) was estimated by simulation at the two experimental pressures, 102 mbar and 200 mbar and an additional reduced pressure of 60 mbar. The initial point (IP) was the labeled alcohol content of each beer 4.7% for G and 5.5% for S. The concentration of ethanol in the beer phase exhibited an exponential-like decay against the vapor fraction (Fig. 3). The % of vapor fractions at their correspondent times in the experiment is shown in Table 1. In general, 1.0% ABV was obtained at about 15% of liquid vaporization. In this study we have analyzed and simulated the compositions considering the instant volume during the process. So, we have not corrected the values considering a possible final dilution with water to the initial volume. This means that if the final residue (dealcoholized beer) would be diluted to the initial volume (e.g., adding water), the % ABV achieved would be lower than 1% of ethanol (that was obtained at 200 mbar for instance). This fact is illustrated in Fig. 4, where we compare the % ABV for diluted distillate and for not diluted distillate. 3.2. Differences of the volatile compound profile during the laboratory scale vacuum distillation process The volatile compound fraction in beer, apart from ethanol, is mainly comprised of higher alcohols formed during primary beer fermentation (Blanco et al., 2014). Higher alcohols contribute to the aroma of beer and produce a warm mouthfeel (Willaert & Nedovic, 2006). The most significant contribution is owed to propanol, isobutanol and isoamyl alcohols (2 and 3-methylbutanol) (Blanco et al., 2014; Brányik, Vicente, Dostálek, & Teixeira, 2008). Higher alcohols are the immediate precursors of most flavor active esters; hence, formation of higher alcohols needs to be controlled to ensure optimal ester production (Gonçalves et al., 2014) because of esters have very low flavor thresholds and a major impact on the overall flavor (Willaert & Nedovic, 2006). However, low content of aroma compounds in alcohol free beers could be attributed to the dealcoholization process (Riu-Aumatell, Miró, Serra-Cayuela, Buxaderas, & López-Tamames, 2014). When we analyzed both regularbeers, results showed (Table 2) that for all volatile compounds the concentrations were higher for the G sample than for the S sample with the exception of 2-methylbutanol, which was higher for the S sample (13.37 mg/L). Calculating the percentage of volatiles (different than ethanol) lost in the dealcoholization process at 102 mbar and 200 mbar at the end of the experiment, it was found that almost all volatile compounds studied were evaporated alongwith ethanol withthe exception of 2-phenyletanol.For the S sample, losses of 97% of esters and 88% of alcohols were observed at 102 mbar and 76% of esters and 95% of alcohols at 200 mbar. For the G sample losses of 96% of esters and 92% of alcohols were achieved at 102mbar, and 90% of esters and 95% of alcohols for 200 mbar. Thesevolatile compound losses can be compared with those reported in previous studies by other authors using different dealcoholization processes (Table 3). From our results, we can concludethat pressure does not have a substantial impact on the relative volatility between the ethanol and other aroma components; therefore, we suggest that the final volatile profile after dealcoholization cannot be significantly improved by only modifying the pressure. Thus, the main alcohols and esters could be affected by the higher temperature applied at 200 mbar. From data in Table 2 it can be seen that at 200 mbar and 67 °C the volatile compounds losses were higher for all compounds except for the amyl alcohols in S sample and ethyl acetate in G sample. Looking at the seven volatile compounds quantified in this study (Table 2), for the ethyl acetate, the evaporation was almost completed at the first 7.53% vapor fraction (Vf), correspondent with the average of % Vf at 15 min of the process (Table 1), in both samples and pressures (from initial values of 17.82 and 26.54 mg/L to 1.07 and 3.65 at 102 mbar; and to 4.09 and 5.18 mg/L at 200 mbar, for the samples S and G respectively), although for the 200 mbar pressure the evaporation seems more gradually. Table 3 Losses of total esters and alcohols in percentage (%) by different alcohol free beer production processes: lab-scale vacuum distillation (this work, present as the average of both samples losses), osmotic distillation (Liguori et al., 2015), vacuum rectification (Montanari et al., 2009), falling film evaporation, dialysis (Liguori et al., 2015) and reverse osmosis (Stein, 1993). Lab-scale vacuum distillation Osmotic distillation Vacuum rectification Falling film evaporation Dialysis Reverse osmosis Total esters 97 (102 mbar) 83 (200 mbar) 99 100 95–100 99 78 Total alcohols 90 (102 mbar) 95 (200 mbar) 77 78 95–98 96 69 Table 4 Boiling points (°C) of the volatile compounds at the different experiment pressures. Boiling points (°C) Compounds Atmospheric pressure 102 mbar 200 mbar Ethyl acetate 77.1 13.7 32.3 1-Propanol 97 33.6 52.2 Isobutanol 107.9 44.5 63.1 Isopentyl acetate 142 78.6 97.2 2-Methylbutanol 127.5 64.1 82.7 3-Methylbutanol 131.1 67.7 86.3 2-Phenyl ethanol 220 156.6 175.2 Table 5 Estimated parameters for Element-1 of Wilson-2 equation in HYSYS. Sucrose Ethanol Ethyl acetate 1-Propanol Isobutanol Isopentyl acetate 2-Methylbutanol 3-Methylbutanol 2-Phenylethanol Water Nitrogen Sucrose 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Ethanol 0.0000 0.0000 2.0000 −12.6000 −1.8000 9.5000 2.5000 2.5000 0.0000 −0.0503 0.0000 Ethyl acetate 0.0000 0.3000 0.0000 2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 −1.8000 0.0000 1-Propanol 0.0000 12.6000 −0.8296 0.0000 0.9130 0.0000 3.0350 3.0350 0.0000 1.1919 0.0000 Isobutanol 0.0000 1.8000 0.0000 −0.7573 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Isopentyl acetate 0.0000 −9.5000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.0000 0.0000 2-Methylbutanol 0.0000 −0.3000 0.0000 −2.0368 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 3-Methylbutanol 0.0000 −0.7256 0.0000 −2.0368 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2-Phenylethanol 0.0000 0 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Water 0.0000 −2.5035 −1.8000 −7.0000 0.0000 −2.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Nitrogen 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 6C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 1-Propanol was almost completely gone within the first of 10.52% Vf in both experiments and beers apart from the S sample at 200 mbar, in which its lost extended to a 14.43% Vf. Isobutanol was evaporated gradually over the process time course under both experimental conditions but, at the first 10.52% Vf more than a half of the concentration was removed (from 9.41 and 10.47 mg/L to 4.68 and 4.32 at 102 mbar; and to 4.25 and 3.46 mg/L at 200 mbar, for thesamples S and G, respectively). A similartrend was observed for isopenthyl acetate, but in this case more than a half was removed during the first 7.53% Vf. For both experiments and samples during the first 7.53% Vf the amount of amyl alcohols (2-methylbutanol and 3-methylbutanol) was reduced by approximately 50%, except for the G sample at 102 mbar. At the end of the laboratory dealcoholization process the amyl alcohols were in higher concentration for the S sample in both experiments (102 mbar, 50 °C and 200 mbar, 67 °C). At the end of both dealcoholization processes (17.87% Vf) the concentrations of the majority of the volatile compounds analyzed were higherfortheSsample. The aromatic alcohol 2-phenylethanol causes ‘sweet’or ‘rose’flavors in beer (Šmogrovičová & Dömény, 1999). Surprisingly, inthis laboratory scale dealcoholization process 2-phenylethanol was apparently produced during the experimental process. This compound has a high boiling point (Table 4), and it was expected to slightly increase its concentration due to the vaporization process (that reduces the volume of the remainingliquid). This feature wassimulated using Aspen HYSYS, obtaining that 2-phenylethanol should have increased its concentration by 3 to 5% maximum, as reported previously by Zürcher et al. (2005). However, its concentration increased after the distillation process by about 30 to 50%, that is from 37.69 ppm up to 59.97 ppm (G at 200 mbar, 67 °C) and75.17 ppm (G at 102 ppm, 50 °C), and from an initial of 34.01 ppm up to 70.65 ppm (S at 200 mbar) and 85.28 ppm (S at 102 mbar). It is well known that, during fermentation, 2-phenyletanol is formed by phenylalanine catabolism (Kobayashi et al., 2008). Higher alcohols achieve maximum concentrations during batch fermentation at a time roughly coincident with cell growth arrestand minimum free amino nitrogen (FAN) concentration. Their formation takes place by the socalled anabolic and catabolic routes. In the anabolic route the 2-oxo acids, arising from carbohydrate metabolism, are decarboxylated to form aldehydes, which are reduced to the corresponding alcohols. Simultaneously, 2-oxo acids also derived from amino acid utilization, which is termed the catabolic (Ehrlich) route to higher alcohol formation. The final concentration of higher alcohols is therefore determined by the uptake efficiency of the corresponding amino acid and the sugar utilization rate. The contribution of each biosynthetic pathway is influenced by the amino acid composition of the wort, fermentation stage and yeast strain. In addition, some higher alcohols may originate from the reduction of aldehydes and ketones that are present in the wort (Brányik et al., 2008). However, in the present study, the beers used were commercial beers, so they were filtered and no fermentation option is possible. Therefore, we explain this effect by the potential degradation and/or transformation of other beer components due to a combined effect of temperature and residence time. It has been shown that, at industrial scale, beer stays only for a few seconds in the dealcoholization processes as it happens in thin film evaporators or spinning cone columns (Brányik et al., 2012). Conversely, in the experimental setup used in this study, the interfacial area of evaporation was considerable lower than that in thin film evaporators, and, hence, the time required for reaching the final ethanol content(≤1%) was calculated to be nearly 45 min, a residence time long enough to give rise to transformation of other potentially related compounds. In this sense, one of the possible formation routes is from the degradation of the amino acid 2-phenylalanine, but any other component from the same metabolic route, e.g., phenyl pyruvate, phenyl acetaldehyde or phenyl acetic acid can additionally lead to 2-phenyethanol in an acidic Table 6 Estimated parameters for Element-2 of Wilson-2 equation in HYSYS. Sucrose Ethanol Ethyl acetate 1-Propanol Isobutanol Isopentyl acetate 2-Methylbutanol 3-Methylbutanol 2-Phenylethanol Water Nitrogen Sucrose 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Ethanol 0.0000 0.0000 −1000.0000 2066.8071 1159.7832 −3062.4265 −500.0000 −500.0000 0.0000 −69.6372 0.0000 Ethyl acetate 0.0000 −300.0000 0.0000 −500.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1-Propanol 0.0000 51.4504 −500.0000 276.3888 −85.8744 0.0000 0.0000 0.0000 0.0000 −132.9251 0.0000 Isobutanol 0.0000 −1675.7465 0.0000 −29.2113 0.0000 0.0000 −442.0000 −442.0855 0.0000 −247.3062 0.0000 Isopentyl acetate 0.0000 3724.3137 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 −1319.7350 0.0000 2-Methylbutanol 0.0000 314.0000 0.0000 843.0000 243.0000 0.0000 0.0000 0.0000 0.0000 −462.0000 0.0000 3-Methylbutanol 0.0000 314.4464 0.0000 843.8578 243.3993 0.0000 0.0000 0.0000 0.0000 −462.4493 0.0000 2-Phenylethanol 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Water 0.0000 346.1512 0.0000 625.5155 −1633.2924 −1716.3821 −2102.0000 −2102.8264 0.0000 0.0000 0.0000 Nitrogen 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 7C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 hydrogen donor bulk liquid (i.e., water/ethanol) such as beer. When a prolonged heating of beer is made, probably the remained content of this amino acid or other similar compound forms the compound by reaction. Accordingly, 2-phenylethanol might be used as a marker of overheating or overtiming in beer dealcoholization processes. 3.3. Simulation results and thermodynamic parameters In order to demonstrate the feasibility of a dynamic Aspen HYSYS simulation for the dealcoholization process, several thermodynamic packages were considered. Inthe present study, it was necessary to consider an EoS with interaction in liquid phase, such as NRTL or Wilson. For our simulation the best results were found using Wilson-2 thermodynamic package from HYSYS database. However, the deviations of the simulation results against the experimental results were unacceptable using the default parameters of the software. Thus, we have performed a fit of the selected binary interaction coefficients for the main measured compounds at 15 min, and then the simulation was tested to check whether the system was able to predict or not the other experimental data points. The best fit parameters for Wilson-2 Element-1 and Element-2 (i.e., interaction parameters according to Aspen HYSYS nomenclature) are listed in Tables 5 and 6 (see also Figs. 5 and 6 for component concentration graphs). The predictions for the seven compounds analyzed were very acceptable. Average absolute deviations (determined as the absolute value of thesimulated instant concentration minus theexperimental instant concentration, divided by the initial value of the concentration) were between 6.9 and 15.1% for both S and G beers (excluding the 0.0 5.0 10.0 15.0 20.0 25.0 0% 5% 10% 15% 20% ppm % vapor fraction Ethyl acetate G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 10.0 0% 5% 10% 15% 20% ppm % vapor fraction Isobutanol G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 0.5 1.0 1.5 2.0 2.5 0% 5% 10% 15% 20% ppm % vapor fraction Isopentyl acetate G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 0% 5% 10% 15% 20% ppm % vapor fraction 2-Methylbutanol G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 10.0 20.0 30.0 40.0 50.0 0% 5% 10% 15% 20% ppm % vapor fraction 3-Methylbutanol G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 0% 5% 10% 15% 20% ppm % vapor fraction 2-Phenylethanol G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 0% 5% 10% 15% 20% ppm % vapor fraction 1-Propanol G SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 Fig. 5. Concentration profiles of the main aroma compounds analyzed in the German beer (G) after the dealcoholization process. 8C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017 values of 2-phenylethanol that behaves oddly). The values obtained by simulation (SIM) and experimentation (EXP) are listed in Table 2 (see also Figs. 5 and 6). For the case of 2-phenylethanol it is clear that the component is generated byreaction, sothe simulation cannot predictit as the assumption 5(seeSection 2.4)isnotfulfilled. Considering the difficulty of the analysis and the system itself we consider the simulation values for prediction acceptable. This is the first time, to our knowledge, that beer is dealcoholized and the experimental values are fit to a simulation and thermodynamic model aimed at creating a prediction tool. From our point of view, the prediction could be improved by studying the kinetics of formation of 2-phenylethanol and by studying a pilot scale plant using a short-residence time equipment (such as falling fill evaporator), but this is out of the scope of this paper. Nevertheless, 2phenylethanol appeared from 15 min on, so this means that the thermodynamic approach is valid for times below that time that indicates that it could be used for simulation of short residence time pieces of equipment. 4. Conclusions Low alcohol and free alcohol beers from thermal dealcoholization (e.g., vacuum distillation) lack of the flavor and aroma compounds that the original beers possesses. Literature data on this is scarce and, so far, no simulation tools to predict the compositions during the dealcoholization process have been published. In this study, we have observed how flavor compounds analyzed vanished to very low concentration levels in the lab-scale vacuum 0.0 5.0 10.0 15.0 20.0 25.0 0% 5% 10% 15% 20% ppm % vapor fraction Ethyl acetate S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 10.0 0% 5% 10% 15% 20% ppm % vapor fraction Isobutanol S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 0.5 1.0 1.5 2.0 2.5 0% 5% 10% 15% 20% ppm % vapor fraction Isopentyl actate S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 0% 5% 10% 15% 20% ppm % vapor fraction 2-Methylbutanol S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 10.0 20.0 30.0 40.0 50.0 0% 5% 10% 15% 20% ppm % vapor fraction 3-Methylbutanol S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 0% 5% 10% 15% 20% ppm % vapor fraction 2-Phenylethanol S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 0.0 2.0 4.0 6.0 8.0 0% 5% 10% 15% 20% ppm % vapor fraction 1-Propanol S SIM 60 SIM 102 EXP 102 SIM 200 EXP 200 Fig. 6. Concentration profiles of the main aroma compounds analyzed in the Spanish beer (S) after the dealcoholization process. 9C. Andrés-Iglesias et al. / Food Research International xxx (2015) xxx–xxx Please cite this article as: Andrés-Iglesias, C., et al., Simulation and flavor compound analysis of dealcoholized beer via one-step vacuum distillation, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.07.017